Why Ex Vivo Occupancy Can Differ From In Vivo Occupancy
An occupancy assay performed on collected tissue does not necessarily report the fraction of targets occupied at the instant the tissue was collected. Collection, dilution, homogenization, washing, and delays before measurement can change the balance between free and target-bound drug. A drug may dissociate after collection, lowering measured occupancy; residual free drug may bind unoccupied targets during processing, raising it. The measured value can therefore reflect both in vivo binding and events in the sample tube.
The central distinction is between occupancy at collection and occupancy at measurement. The first is the biological quantity of interest. The second is what the assay encounters after the sample has undergone its particular handling history. A defensible interpretation requires describing that history and evaluating whether it could have changed target binding.
This distinction matters even when the assay itself is technically precise. A precise measurement of \(\theta(t)\) is not automatically an accurate estimate of \(\theta_0\). The earlier material on fractional target occupancy and binding kinetics provides the relevant equilibrium and rate concepts; here the key issue is how collection changes the conditions under which those processes continue.
A Kinetic Model for Processing
Consider a reversible interaction between drug and target. Let \(C\) be the free drug concentration during processing, \(k_{\mathrm{on}}\) the association rate constant, and \(k_{\mathrm{off}}\) the dissociation rate constant. Assume target abundance is unchanged over the short processing interval, binding sites are equivalent, and the measured occupancy corresponds to the fraction of those sites occupied. If \(C\) is approximately constant, mass-action kinetics give the occupancy change as association onto unoccupied sites minus dissociation from occupied sites.
This equation can be rearranged into a linear first-order differential equation. Its equilibrium occupancy at the processing concentration is \(\theta_{\mathrm{eq}}=k_{\mathrm{on}}C/(k_{\mathrm{on}}C+k_{\mathrm{off}})\), and the rate at which processing occupancy approaches that equilibrium is \(k_{\mathrm{on}}C+k_{\mathrm{off}}\).
The equation is conditional on the assumptions above: in particular, constant free concentration and no relevant target loss or alteration. In real samples, free concentration may change as drug is diluted, released from tissue compartments, or bound to other material. The equation is then a useful local or simplified model, not a guarantee that the sample follows a single constant-\(C\) trajectory.
Dilution and Dissociation Under Sink Conditions
Dilution reduces free drug concentration. If the post-collection concentration is sufficiently low that association is negligible, the processing sample is approximately under sink conditions: dissociated drug is diluted or removed, and little rebinds. Mathematically, \(k_{\mathrm{on}}C\) is negligible compared with \(k_{\mathrm{off}}\), so the kinetic equation becomes \(d\theta/dt\approx-k_{\mathrm{off}}\theta\). Separating variables and integrating from collection time \(0\) to processing time \(t\) gives the exponential loss relationship:
Thus, even if occupancy is high at collection, a sufficiently long delay can reduce measured occupancy for a rapidly dissociating drug. A large dilution factor alone does not specify the size of the loss: the outcome also depends on dissociation kinetics and the time spent under the diluted conditions. Conversely, a slow dissociation rate can preserve occupancy during a short processing interval despite substantial dilution.
Worked Example: Occupancy Lost During Dilution and Delay
Suppose a hypothetical tissue sample has \(\theta_0=0.80\) at collection. After a large dilution, assume free drug is negligible, rebinding does not occur, and the drug dissociates with \(k_{\mathrm{off}}=0.080\ \mathrm{min}^{-1}\). The sample is processed for 5.0 minutes before occupancy is measured. Under the sink-condition model:
Since \(e^{-0.400}=0.6703\), the predicted measured occupancy is \(0.80(0.6703)=0.5362\), or about 53.6%. The hypothetical sample began at 80.0% occupancy, so the model predicts a loss of 26.4 percentage points during processing. This calculation is appropriate only if free drug remains low enough that rebinding is negligible and the dissociation rate is applicable under the processing conditions.
Residual Drug and Ex Vivo Rebinding
Dilution does not always create sink conditions. Free drug may remain in the sample, and drug released from tissue or dissociated from target may also become available to bind again. If the processing concentration is approximately constant and appreciable, occupancy can move toward \(\theta_{\mathrm{eq}}\) rather than simply decay. When \(C\) is constant, the equilibrium expression can also be written using \(K_D=k_{\mathrm{off}}/k_{\mathrm{on}}\):
Here the formula describes equilibrium with the processing concentration, not necessarily equilibrium in the living subject. If processing concentration is above the in vivo free concentration, ex vivo occupancy may rise. If it is lower, occupancy may fall. Whether the assay sees such a change depends on processing duration relative to the kinetic rate \(k_{\mathrm{on}}C+k_{\mathrm{off}}\).
Worked Example: Residual Drug Raises Occupancy During Processing
Consider a hypothetical compound with \(k_{\mathrm{on}}=0.020\ \mathrm{nM}^{-1}\mathrm{min}^{-1}\) and \(k_{\mathrm{off}}=0.100\ \mathrm{min}^{-1}\). Suppose the processed sample retains a constant free concentration of \(C=10.0\ \mathrm{nM}\), and occupancy at collection is \(\theta_0=0.200\). The implied \(K_D\) is \(k_{\mathrm{off}}/k_{\mathrm{on}}=5.00\ \mathrm{nM}\). Therefore, the processing equilibrium is:
The combined rate is \(k_{\mathrm{on}}C+k_{\mathrm{off}}=0.300\ \mathrm{min}^{-1}\). After 4.0 minutes, the predicted occupancy is:
Because \(e^{-1.20}=0.3010\), \(\theta(4.0)=0.6667-0.1405=0.5262\), or about 52.6%. Occupancy has risen from 20.0% at collection to 52.6% in the modeled sample. This is not evidence of increased in vivo binding; it is the predicted approach toward equilibrium at the residual processing concentration. The result relies on that concentration remaining constant, which may be questionable if the sample is diluted further or drug is depleted by binding.
Controls That Separate In Vivo Binding From Processing Effects
No single control proves that a measured occupancy value is artifact-free. A useful strategy combines controls that expose time dependence, concentration dependence, and the capacity of post-collection drug to bind target. The controls should match the tissue, temperature, dilution, and handling steps used for the experimental samples as closely as possible.
Divide matched samples and process them with different collection-to-measurement delays. A systematic change with delay indicates that occupancy is not stable throughout handling. Under sink conditions, an approximately linear relationship between \(\ln\theta(t)\) and time supports the simple dissociation model.
Compare handling conditions that produce different free-drug concentrations. A change in measured occupancy with dilution or wash conditions is consistent with concentration-dependent rebinding or dilution-driven dissociation. Confirm that the manipulation itself does not damage target or change assay recovery.
Add drug to a drug-naive tissue sample at collection or during processing, then apply the same procedure. Detectable occupancy in this control shows that drug present after collection can occupy target ex vivo under those conditions.
Use a rapid quench or other validated means of arresting binding changes, alongside the standard procedure. Agreement is reassuring; disagreement signals sensitivity to handling. The preservation method must itself be checked because it may disrupt complexes or alter the assay signal.
A processing time series can also help estimate the occupancy at collection, but only if its model assumptions are credible. Under sink conditions, two or more time points may support an exponential back-extrapolation. If residual concentration causes rebinding, a sink-condition extrapolation will generally be wrong; one must instead characterize the processing concentration and binding kinetics or avoid relying on a model correction.
Worked Example: Back-Extrapolating a Sink-Condition Time Series
In a hypothetical processing study, matched samples yield occupancies of 0.720 at 1.0 minute and 0.550 at 4.0 minutes. Assume target is stable, free drug is negligible throughout, and dissociation follows a single exponential with one \(k_{\mathrm{off}}\). The ratio of the two measurements removes the unknown collection occupancy:
Since \(0.550/0.720=0.7639\), \(\ln(0.7639)\approx-0.2691\), and the estimated dissociation rate is \(0.0897\ \mathrm{min}^{-1}\). Using the 1.0-minute measurement, the estimated collection occupancy is:
The model therefore estimates occupancy at collection as about 78.8%. This value is a model-based extrapolation, not a direct observation. It should not be treated as reliable if residual drug permits rebinding, if the rate changes during processing, or if the two samples differ for reasons other than delay.
Diagnostics, Alternatives, and Limitations
Several patterns warn that processing is influencing the result. Occupancy that changes with delay, dilution, temperature, or washing is handling-sensitive. A drug-naive sample that becomes occupied after being spiked at collection demonstrates the possibility of ex vivo binding. Non-exponential time courses, or different apparent rates under ostensibly identical conditions, challenge the simple one-state model. They may reflect changing free concentration, multiple target populations, tissue compartments, or changes in target integrity.
The main alternatives are procedural rather than interchangeable statistical corrections. Rapid measurement or preservation aims to minimize the time available for binding changes, but the preservation step may alter the target or assay. Excessive dilution or washing can reduce rebinding, but may accelerate dissociation and thereby lower occupancy. Time-course modeling can estimate the collection-time value, but depends on the assumed kinetic model and reliable processing conditions. These approaches answer different practical needs; a control demonstrating stability under the chosen handling procedure is often more persuasive than an unvalidated correction.
A common mistake is to assume that extensive washing necessarily preserves the in vivo occupancy because it removes unbound drug. Washing can indeed reduce rebinding, but it also lowers free concentration and can pull the reversible binding system toward dissociation. The correction is to assess both effects: measure occupancy across processing delays or wash conditions, and interpret any correction using the relevant kinetics rather than treating a washed value as automatically equal to \(\theta_0\).
The simple equations have important limitations. They assume a well-defined target pool and either negligible or constant free concentration; tissue samples may violate both conditions. Drug can partition between compartments, bind nonspecifically, or be released gradually. Target may be degraded or become inaccessible, and the assay may report a signal that is not a direct occupancy fraction. Such complications require evidence and a more appropriate model; a fitted exponential alone cannot establish that the model is biologically correct. The practical goal is to make the collection-to-measurement process short and reproducible, then demonstrate with controls that residual processing effects are small or characterize them transparently.
Check Your Understanding
Answer each question using the stated assumptions. Brief solutions follow the questions.
- Under sink conditions, a sample has occupancy 0.60 at collection and \(k_{\mathrm{off}}=0.050\ \mathrm{min}^{-1}\). What occupancy does the model predict after 6.0 minutes?
- Why can washing reduce an ex vivo occupancy artifact and nevertheless cause measured occupancy to fall?
- A drug-naive tissue sample is spiked with drug at collection and later shows target occupancy. What does this control establish, and what does it not establish?
- In the constant-concentration model, what happens to occupancy over time if \(\theta_0\) is below \(\theta_{\mathrm{eq}}\)?
Solutions. (1) \(\theta(6.0)=0.60e^{-(0.050)(6.0)}=0.60e^{-0.300}=0.4445\), or about 44.5%. (2) Washing can remove free drug and limit rebinding, but the resulting low concentration also permits bound drug to dissociate without replacement. (3) It establishes that post-collection drug can bind target under the tested handling conditions; it does not show that the treated subject’s measured occupancy was entirely produced ex vivo or quantify that contribution without further controls. (4) The solution \(\theta(t)=\theta_{\mathrm{eq}}+(\theta_0-\theta_{\mathrm{eq}})e^{-(k_{\mathrm{on}}C+k_{\mathrm{off}})t}\) approaches \(\theta_{\mathrm{eq}}\) from below.